Did you follow through on Ian's suggestion and short U2 grid, and then go back one stage and short U2A. I'm guessing the answer to both those is an undetectable hiss from the speaker, or if you have a sensitive AC wideband voltmeter then level is down with the voltmeter noise, or your soundcard setup shows little change from shorted soundcard input (do this with soundcard ground connection still made to amp, as ground loop signals can easily mislead).
Ciao, Tim
Ciao, Tim
Therein lies your problem. Even if this was a perfect mic preamp its EIN could not be better than -130dBu which means with 65dB of gain its output noise is going to be no lower than -65dB (130 -65).
I think you said when you shorted the input there was 60K or so still in series. A 60K resistor has a Johnson noise of -107dBu which after 65dB of gain becomes -42dBu at the output. No wonder you can hear it at 4ft.
I think the question you have to ask yourself is do you really need this much gain?
Cheers
Ian
Could be a bad part. I'd go through the circuit, output to input, shorting nodes to ground anywhere it's safe to do that (perhaps through a big cap so DC levels aren't upset). I'd also replace that bias LED in the front end tube with a resistor either way though. Any noise or distortion that LED produces will get amplified by every stage in the entire amp. Don't pretend it's not in the signal path.
Sorry I'm a little slow on responding, I've got to get ready for a funeral.
I tried a 750R resistor wit the bypass cap and it made no noticable difference.
Here video I uploaded to youtube from a Nikon Coolpix. It seems a bit noisier than the speaker is.
The speaker is a 1976 Celestion G12M. I think the efficiency should be about the same as the Heritage Greenback which is 96dB/W-M.
Recording is done with the recorder 1M from the speaker. The tone is 1KHz at 6mV when measured on a 8 ohm load.
I'll try to get to shorting input and grids tomorrow per Ian's suggestion.
YouTube - March 7 2011 001
I tried a 750R resistor wit the bypass cap and it made no noticable difference.
Here video I uploaded to youtube from a Nikon Coolpix. It seems a bit noisier than the speaker is.
The speaker is a 1976 Celestion G12M. I think the efficiency should be about the same as the Heritage Greenback which is 96dB/W-M.
Recording is done with the recorder 1M from the speaker. The tone is 1KHz at 6mV when measured on a 8 ohm load.
I'll try to get to shorting input and grids tomorrow per Ian's suggestion.
YouTube - March 7 2011 001
Yes a guitar amp in the style of a High Gain Lead/overdrive amp, modeled after the TrainWreck Liverpool with the changes of;
(1) replacing all 12AX7 stages with 6N2P-EV stages
(2) reduce the output from 4 6BQ5/EL84s Parallel push pull to two 6P1P (6AQ5 equivalents ?) push pull, including output transformer change
(3) convert an un-used triode to cathode follower between first gain stage and tone stack
(4) several tweaks that reduced the gain slightly in an attempt to reduce hiss
(1) replacing all 12AX7 stages with 6N2P-EV stages
(2) reduce the output from 4 6BQ5/EL84s Parallel push pull to two 6P1P (6AQ5 equivalents ?) push pull, including output transformer change
(3) convert an un-used triode to cathode follower between first gain stage and tone stack
(4) several tweaks that reduced the gain slightly in an attempt to reduce hiss
I finally got back to this. Some times it is best to set a problem aside and let your mind clear to you can take a fresh look at it.
I put the amp on the bench and hooked it up to AudioTester and ran a frequency sweep tonight. No wonder it had so much hiss. With over 20KHz bandwidth and that kind of gain it is bound to have hiss.
I found I had placed two cap values wrong. One effected the low frequency roll off which was not the issue. The second one effected the hf roll off. Changing it made a world of difference. The amp does not have quite as much gain but is so much better sounding.
The first plot shows the original frequency response compared to the corrected response.
The second plot shows all controls at noon, Min base w/ Max treble, and max base w/ min treble.
Driving a "1976 Greenback" with the volume at max I have to put my ear to the speaker to hear the hiss.
I put the amp on the bench and hooked it up to AudioTester and ran a frequency sweep tonight. No wonder it had so much hiss. With over 20KHz bandwidth and that kind of gain it is bound to have hiss.
I found I had placed two cap values wrong. One effected the low frequency roll off which was not the issue. The second one effected the hf roll off. Changing it made a world of difference. The amp does not have quite as much gain but is so much better sounding.
The first plot shows the original frequency response compared to the corrected response.
The second plot shows all controls at noon, Min base w/ Max treble, and max base w/ min treble.
Driving a "1976 Greenback" with the volume at max I have to put my ear to the speaker to hear the hiss.
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.